Ranking Policy Gradient

Authors: Kaixiang Lin, Jiayu Zhou

ICLR 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct extensive experiments showing that when consolidating with the off-policy learning framework, RPG substantially reduces the sample complexity, comparing to the state-of-the-art.
Researcher Affiliation Academia Kaixiang Lin Department of Computer Science and Engineering Michigan State University East Lansing, MI 48824-4403, USA linkaixi@msu.edu Jiayu Zhou Department of Computer Science and Engineering Michigan State University East Lansing, MI 48824-4403, USA jiayuz@msu.edu
Pseudocode Yes Algorithm 1 Off-Policy Learning for Ranking Policy Gradient (RPG)
Open Source Code Yes Code is available at https://github.com/illidanlab/rpg.
Open Datasets Yes To evaluate the sample-efficiency of Ranking Policy Gradient (RPG), we focus on Atari 2600 games in Open AI gym Bellemare et al. (2013); Brockman et al. (2016)
Dataset Splits No The paper does not explicitly provide training, validation, and test dataset splits with percentages or sample counts.
Hardware Specification No The paper does not provide specific details about the hardware used to run the experiments, such as GPU or CPU models.
Software Dependencies No The paper mentions "Dopamine framework" and "openai baselines" but does not specify their version numbers or other software dependencies with version numbers.
Experiment Setup Yes The network architecture is the same as the convolution neural network used in DQN Mnih et al. (2015). We update the RPG network every four timesteps with a minibatch of size 32. The replay ratio is equal to eight for all baselines and RPG (except for ACER we use the default setting in openai baselines Dhariwal et al. (2017) for better performance). ... Table 3: Hyperparameters of RPG network (Hyperparameters Value Architecture Conv(32-8 8-4) -Conv(64-4 4-2) -Conv(64-3 3-2) -FC(512) Learning rate 0.0000625 Batch size 32 Update period 4 Margin in Eq (6) 1)